🥟 Chao-Down #160 Google introduces a watermark that can help detect AI-generated images, General Motors brings Google's chatbot to cars, Several top brands have opted out of OpenAI's web crawler

Google Cloud made their flurry of product announcements, staking their name in the ground with Duet AI for Google Workspace. Charged at $30 per month which is the same as Microsoft’s 365 Copilot, Google believes it can win the enterprise with these new AI offerings.

But the one that caught my attention the most was a new product from Deepmind called SynthID, a watermark for images that is supposed to be impossible to see for a human but easy for a detection tool to spot.

AI-generated images are becoming more and more pervasive and we’ll need more tools like this in our arsenal to make sure we can navigate an increasingly AI-driven world.

All in all, exciting work from Google!

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

Investors bet on generative AI to revolutionize drug discovery (PitchBook)

Google Will Let You Send A Bot To A Meeting On Your Behalf (Forbes)

The SynthID watermark from Google DeepMind can detect AI-generated images - (The Verge)

GM is using Google’s AI chatbot to handle simple OnStar calls (The Verge)

Dozens of big brands have blocked GPTBot, OpenAI's new web crawler (searchengineland.com)

ABA taps prominent lawyers to tackle AI risks, opportunities (Reuters)

Always be Learnin’ 📕 📖

MapReduce, TensorFlow, Vertex: Google's bet to avoid repeating history in AI (supervised.news)

Llama 2 is about as factually accurate as GPT-4 for summaries and is 30X cheaper | (Anyscale)

Measuring developer productivity? A response to McKinsey (pragmaticengineer.com)

Projects to Keep an Eye On 🛠

getumbrel/llama-gpt: A self-hosted, offline, ChatGPT-like chatbot. Powered by Llama 2. 100% private, with no data leaving your device. New: Code Llama support! (Github)

embedchain/embedchain: Framework to easily create LLM powered bots over any dataset. (Github)

dot-agent/openagent: Microservices approach to AGI. Modular components for AI apps or AGI agents. (... and solving some wicked LLM problems like ⚡ 2X faster LLaMA 2) (Github)

The Latest in AI Research 💡

Robust Distortion-free Watermarks for Language Models (arxiv)

Evaluating progress in automatic chest X-ray radiology report generation: Patterns (cell.com)

AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework (arxiv.org)

The World Outside of AI 🌎

You’ll Probably Never Know How Much Protein You Actually Need (The Atlantic)

Covid Closed the Nation’s Schools. Cleaner Air Can Keep Them Open. (The New York Times)

Everyone on LinkedIn is absolutely crushing it – or so it seems | US small business | The Guardian

My Family Spent $11,000 Trying to Rid Our House of Lead (The Atlantic)

Magic Mushrooms. LSD. Ketamine. The Drugs That Power Silicon Valley. (WSJ)

One Last Bite 😋

One model to rule them all for full self-driving? Elon thinks so!